Skip to content

feat(ai): Tool call loop, built-in data tools, agent runtime & chat API #1016

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
hotlong merged 5 commits into from
Mar 31, 2026
Merged

feat(ai): Tool call loop, built-in data tools, agent runtime & chat API #1016

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
1838110
Initial plan
Copilot Mar 31, 2026
71909ed
feat(ai): implement tool call loop, data tools, agent runtime, and ag…
Copilot Mar 31, 2026
f4e8229
test: add 42 tests for chatbot features and export Agent type from spec
Copilot Mar 31, 2026
23b5335
fix: address code review feedback — fix maxIterations logic, remove a…
Copilot Mar 31, 2026
e68ef7d
fix: address all 7 PR review comments
Copilot Mar 31, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 23 additions & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,29 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
## [Unreleased]

### Added
- **`@objectstack/service-ai` — Data Chatbot: Tool Call Loop & Agent Runtime** — Implements
an Airtable Copilot-style data conversation Chatbot with full-stack support:
- `AIService.chatWithTools()` — automatic multi-round LLM ↔ tool call loop with
`maxIterations` safety limit, parallel tool execution, and forced final response
- `AIResult.toolCalls` — new field on the AI result contract so adapters can return
tool call requests from the LLM
- `ChatWithToolsOptions` — new contract interface extending `AIRequestOptions`
- 5 built-in data tools: `list_objects`, `describe_object`, `query_records`,
`get_record`, `aggregate_data` — with parameter schemas, limit capping (max 200),
and error handling
- `registerDataTools(registry, context)` — factory to register all data tools
against `IDataEngine` + `IMetadataService`
- `AgentRuntime` — loads agent metadata, builds system prompts from instructions +
UI context (`objectName`, `recordId`, `viewName`), resolves agent tool references
against the `ToolRegistry`
- `buildAgentRoutes()` — new `POST /api/v1/ai/agents/:agentName/chat` route with
agent lookup, active-check, context injection, and `chatWithTools` integration
- `DATA_CHAT_AGENT` — built-in `data_chat` agent spec with role, instructions,
guardrails, planning config, and tool declarations
- `AIServicePlugin` auto-registers data tools and `data_chat` agent when
`IDataEngine` + `IMetadataService` are available in the kernel
- 42 new test cases covering tool call loop, data tools, agent runtime, agent
routes, and agent spec validation
- **`@objectstack/service-ai` — ObjectQL-backed persistent ConversationService** — New
`ObjectQLConversationService` implements `IAIConversationService` using `IDataEngine`
for durable conversation and message storage across service restarts:
Expand Down
Loading